Architecting, installing and maintaining your SAS environment

SAS references errors after cloning sas servers

Reply
Occasional Contributor
Posts: 8

SAS references errors after cloning sas servers

Hi,

We are working on 2 servers(developer enviroment) that were cloned from a production enviroment, we have made several changes like changing password and reference using sas deployment manager but, we still have errors and references that still point to the production enviroment server, manually we,ve changed some files but we still got errors with websphere enteprise applications, and to use PM via the web app and also we can't use FM.

 

We apply all the hotfixes needed for this servers, thinking that we could solve the problem, but the same errors stil persist.

 

With webpshere we try updating and uninstall/install the enterprise apps, using redeployed .ear files, but the apps do not run yet.

 

Can someone that has worked on cloned servers, can give us some tips about this issues?

Super User
Posts: 3,102

Re: SAS references errors after cloning sas servers

[ Edited ]

In my experience, SAS do not generally recommend cloning SAS servers for the very issues you are running into. Every recent SAS install I've been involved in has been done with separate installs. For example we have a Disaster Recovery mirror of our Production SAS site but that was done as a separate install. One of the big issues is the fact that server host names are hard-coded in many locations in the SAS configuration files.

 

Having said that there is some documentation you could research to see if it can help in your situation. This relates to SAS 9.4:

 

http://support.sas.com/documentation/cdl/en/bisag/68240/HTML/default/viewer.htm#n1ov3paky6afj8n1bj7l...

 

 

 

Valued Guide
Posts: 3,208

Re: SAS references errors after cloning sas servers

Cloning is a normal requirement as of DR plans and BCM. It is a frustrating situation SAS is that neglectant on these type of requiements.  For example: http://www.tldp.org/LDP/Linux-Filesystem-Hierarchy/html/c23.html This Linux / Unix  hierarchy has the goal of a clear functionality segregation. It would make more sense SAS would follow those kind of guidelines instead of fighting them.

 

That said you can place the SASHOME at any location you are needing. It is the same appraoch as having a peacefull coexistance of installations http://support.sas.com/rnd/migration/planning/software/bi.html  By isolating the SASHOME (installation) from the Configuration (ports metadatabase - webcontent). You can clone the SASHOME (installation) without too much issues.
This will accomplish the requirment you test/verify the same thing in UAT before going in Prod.

That will leave you we the WEB applications as they are not maintaianed at OS level but very different the SAS metadata and more.
Not really problematic as you change those after cloning as SASkiwi explained.

The content in metadata and web (Webdav) of course you have a tailored approach for that supporting all normal fucntionilty ot be able to recover anything to an isolated artifact in any time-frames. "Give me the restore of that report of two week ago"

A little bit nasty formulated as I know there is no standard approach for that documented by SAS. 

---->-- ja karman --<-----
Ask a Question
Discussion stats
  • 2 replies
  • 321 views
  • 0 likes
  • 3 in conversation